Surface microbes in the neonatal intensive care unit: changes with routine cleaning and over time

Abstract

Premature infants in neonatal intensive care units (NICUs) are highly susceptible to infection due to the immaturity of their immune systems, and nosocomial infections are a significant risk factor for death and poor neurodevelopmental outcome in this population. To investigate the impact of cleaning within a NICU, a high-throughput short-amplicon-sequencing approach was used to profile bacterial and fungal surface communities before and after cleaning. Intensive cleaning of surfaces in contact with neonates decreased the total bacterial load and the percentage of Streptococcus species with similar trends for total fungal load and Staphylococcus species; this may have clinical relevance since staphylococci and streptococci are the most common causes of nosocomial NICU infections. Surfaces generally had low levels of other taxa containing species that commonly cause nosocomial infections (e.g., Enterobacteriaceae) that were not significantly altered with cleaning. Several opportunistic yeasts were detected in the NICU environment, demonstrating that these NICU surfaces represent a potential vector for spreading fungal pathogens. These results underline the importance of routine cleaning as a means of managing the microbial ecosystem of NICUs and of future opportunities to minimize exposures of vulnerable neonates to potential pathogens and to use amplicon-sequencing tools for microbial surveillance and hygienic testing in hospital environments.

Fig 1

Key to surface swab samples. The neonate-associated surfaces (top) were sampled before and after intensive cleaning. Disposable items (tubing, nose pieces, pacifiers, and suction catheters) were sampled when new and after use. Environmental surfaces of the room (bottom) were sampled four times at weekly intervals, unrelated to the timing of routine cleaning.

Fig 2

Microbial heat maps reveal bacterial distribution in the NICU. Heat maps of select, most abundant bacterial genera detected across each surface type using 16S rRNA gene amplicon sequencing are shown. The keys for each plot indicate the relative abundance scale from blue (absent/low) to red (high). Sites that appear white had bacterial DNA below the limit of detection for this assay.

Fig 3

Microbial heat maps reveal fungal distribution in the NICU. Heat maps of select, most abundant fungal genera detected across each surface type using ITS amplicon sequencing are shown. The keys for each plot indicate the relative abundance scale from blue (absent/low) to red (high). Sites that appear white had fungal DNA below the limit of detection for this assay.

Fig 4

Cleaning drives shifts in microbial structures and abundances. (A) A weighted UniFrac PCoA biplot demonstrates the abundance-weighted clustering of clinical surfaces before cleaning (red) and after cleaning (blue), disposable items before (purple) and after use (yellow), and nonclinical surfaces in the room environment (orange). Each colored orb represents a single sample, and the proximity indicates the degree of similarity between individual samples. In this biplot, bacterial taxa are coplotted as gray orbs to indicate the association with different sample clusters; orb size is a function of the mean abundance. The P value in the top right corner indicates the result of a permutational MANOVA test comparing the bacterial community similarity of pre-/postcleaning and environmental sample groups. (B and C) qPCRs indicate drops in bacterial (B) and fungal (C) surface bioloads in response to cleaning. The error bars represent the standard deviation (SD); P values indicate the results of paired t tests. (D and E) Quantitative heat maps of clinical surfaces illustrate changes in bacterial (D) and fungal (E) bioloads on each surface, as determined by qPCRs. The keys for each plot indicate the absolute abundance scale from blue (absent/low) to red (high). Sites that appear white had DNA below the limit of detection for this assay.